KR20030011919A - Optical inspection method and apparatus with adaptive spatial filter - Google Patents

Abstract

Provided are a method and apparatus for optical inspection of patterned articles. The area on the product is dimmed with the incident beam to form light returned from the dimmed area. An image of the illuminated area is obtained and analyzed to determine the intensity distribution of the scattered light component from the pattern of the illuminated area within a given focusing angle field located outside the solid angle of specularly reflected light propagation. Based on the determined distribution, light components scattered from the dimming area and propagating to at least one predetermined solid angle segment of the predetermined concentration angle are focused and directed to the dark field detection unit.

Description

Optical inspection method and optical inspection device having an adaptive spatial filter {OPTICAL INSPECTION METHOD AND APPARATUS WITH ADAPTIVE SPATIAL FILTER}

Conventional structures of patterned articles, such as semiconductor wafers, include elementary cell elements repeated several times the dimensions of both sides, in which a two-dimensional pattern is generated almost completely periodically. Since timely abnormal shape detection on the surface of the wafer is a very important factor, the yield is increased by inspecting the semiconductor wafer before and after the patterning process.

Inspection of the wafer prior to patterning is based on the fact that light is mainly scattered from the anomaly present on the flat and flat surface of the unpatterned wafer. That is, detection of scattered light may indicate a defect. Applying optical inspection to a patterned wafer for the purpose of detecting defects (eg, the presence of type characters) can cause the scattered light to generate light scattered by the pattern. Thus, the detection of scattered light does not necessarily indicate a defect. Conventional methods of detecting defects are the so-called "die-to-database" and "die-to-die" techniques, whereby scattered from individual dies The light is compared with a previously prepared database representing light scattered from the "defective" die and the "neighbors" of its individual die. The difference between the signals is represented by light scattered from the abnormal shape present on the surface of each product. In general, differences detected in light components scattered from individual dies are considered defective by indicating a lack or addition of some features of the die, as compared to a "defective" die or a "neighbor" die.

Typically, the optical inspection device consists of major structural parts, such as a light control system and a light condensing / detecting system using both bright field or dark field modes. This bright field detection mode is based on the variation in the specular reflectance from the wafer at the time of inspection, formed by defects scattered on the wafer. Dark field detection mode uses scattering from defects out of orientation in specular reflectance.

The general purpose of the condensing scheme is to increase the signal to noise ratio of the detection signal as much as possible. Various kinds of dark field detection schemes are known to be very effective for the purpose of defect detection. According to one known technique of this kind, which facilitates meaningful signal comparisons, disclosed in U.S. Patent Nos. 4,898,471 and 5,604,585, the light collecting system is one at a different azimuth and altitude than the position at which specular reflection occurs. Condenses at a constant condensing angle. However, the scattered light from the patterned product (eg wafer) always includes light components scattered at a condensing angle from the pattern, and therefore detection of such light components may result in "noise" in the detected signal. It appears to be an increase.

FIELD OF THE INVENTION The present invention relates to the field of optical inspection techniques, and relates to methods and apparatus for inspecting patterned products, such as integrated circuits, printed circuit boards, photolithographic masks, liquid crystal displays, and the like using condensing angle design.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments, which are not limited by the embodiments, will be described with reference to the accompanying drawings in order to understand the invention and to see how the embodiments can be performed.

1 schematically shows a main part of an optical inspection device according to an embodiment of the present invention.

2 shows an optical inspection device according to another embodiment of the present invention.

3 schematically shows the main parts of the detection unit of the apparatus of FIG. 1.

4 illustrates a structure of a light collecting device of the detection unit of FIG. 2.

5 illustrates an embodiment of a mask assembly in which the filter of the apparatus of FIG. 1 is suitably used.

6A and 6B respectively show the actual bright field image of the illuminated area and its “ideal” simulation image.

7A-7E show simulation images of each of five different pattern structures.

8 schematically illustrates the dimming action projected on a scattered plane, which is used to simulate a Gaussian beam spanning the surface of the article.

9 schematically illustrates the shape of the scattering problem used for simulation purposes.

10A and 10B show images of simulation scattering results and intensity profiles, respectively.

It is an object of the present invention to improve automatic optical inspection of patterned articles by providing novel methods and apparatus that can significantly increase the signal-to-noise ratio of a detected signal.

The present invention has the advantage of the technique using the variable angle design of the light collecting system. This technology is also disclosed in a concurrent ongoing application that is assigned to the assignee of the present application. According to this technique, a product (e.g., wafer) is scanned per area during inspection, and light scattered from each of these scan areas is collected at a predetermined maximum condensing angle constant for each scan area and detected through a filter. Guided by means. The latter can be picked up from a portion of the totally focused light propagating into the solid angle segment of the total focusing angle, where the intensity of light scattered from the pattern is minimal compared to other solid angle segments of the maximum focusing angle. Is selected. In contrast, the signal-to-noise ratio of the detected signal can be increased. This detecting means comprises at least one detection unit operating in the dark field imaging mode, ie collecting the light collecting components scattered from the product at a different orientation and altitude than where most other specular reflection occurs.

The main idea of the present invention consists in selecting a variable angle solution. This is provided by obtaining a bright field or high resolution dark field image and analyzing the wet image to determine the solid angle segment of the light propagation back from the pattern in the condensing angle field. Thereby, applying a customized light collection (CLC) to the dark field scattering signal prevents the portion of the scattered signal from reaching the detector that is associated with the pattern on the product and constitutes "noise". Only part of the scattering signal coupled to the feature in the region can be detected. That is, the appropriate mask scattered from the product and placed in the optical path of light propagates to the dark field detector, whereby the mask cuts off the solid-angle segment of the propagation of light scattered from the pattern. This significantly increases the signal-to-noise ratio of the detection signal.

Thus, according to a first aspect of the invention, there is provided a method for optical inspection of a patterned article, which method comprises:

(Iii) dimming an area on the product with incident light to generate light returned from the dimming area;

(Ii) acquiring an image of the dimming area and generating data indicative thereof;

(Iii) analyzing the generated data and determining the intensity distribution of the scattered light component from the pattern of the dimming area within a predetermined condensing angle field outside the solid angle of propagation of specularly reflected light; And

(Iii) based on the determined distribution, by condensing the light collected in the predetermined condensing angle field, condensing light components scattered from the dimming area and propagating to at least one predetermined solid angle segment of the predetermined condensing angle field, and condensing Directing the light components to the detection device.

The term "condensing angle field" as used herein denotes the maximum condensing solid angle of light scattered from the product defined by the optical instrument of the detection unit.

In the above-mentioned step (ii), the light detected to obtain an image of the dimming area is either specularly reflected from the dimming area or propagates at a solid angle outside the propagating solid angle of the specularly reflected light (for example, the dark field detection mode). It may also be scattered light. In order to obtain data representing the image obtained in the pattern of the illuminated area, a sufficiently high resolution must be provided. For this purpose, a large number of aperture objective lenses are used for the condensing optical apparatus.

The analysis of the generated data is configured to obtain discontinuous two-dimensional (or three-dimensional) arrays representing so-called "modeling" of the pattern structure and scattered patterns. This data is used to simulate dark-field scattering patterns in the condensing angle field. The simulation results determine the "background" intensity rising from the periodic pattern as the intensity lobe of the simulation / imaging plot, and allow the light component to propagate to solid segments of the condensing angle field different from that of the "background" light component. . This is realized by placing a suitable mask (filter) at a suitable position in the optical path of total condensed light (so-called fourier filtering).

It is to be understood that each predetermined pattern structure is characterized by propagating solid segments of "background" light components scattered from the periodic pattern. Thus, a library (database) can be previously designed to include a plurality of data records, for a plurality of different pattern structures, each represented by a solid segment of propagation of light scattered from a predetermined periodic pattern. These other pattern structures may be manufactured at different manufacturing steps from a given product (eg, wafer), or may be combined with a plurality of different product types. Indeed, such a library represents reference data used to focus light components scattered from the illuminated area into at least one predetermined solid angle segment of a predetermined focusing angle field.

Next, according to another aspect of the present invention, there is provided a method of inspecting a pattern structure at a predetermined maximum solid angle of condensation of light scattered from an inspection target product, wherein the predetermined maximum condensed solid angle is propagated by specularly reflected light specularly reflected from the pattern structure. Is located outside the solid angle of,

Providing a database comprising data indicative of a plurality of solid angle segments of light propagation within a predetermined maximum condensing solid angle, wherein the plurality of solid angle segments correspond to a plurality of pattern structures including a pattern structure to be examined; Each corresponding to a propagation direction of the light component scattered from the corresponding pattern structure, wherein the distribution of light scattered from the pattern is substantially smaller than other solid segments of the solid angle of condensing; And

Selecting data indicative of the pattern structure from a database, the data being used to filter and detect light components propagating from the total light collected at a predetermined maximum condensing angle to the corresponding solid angle segments; .

Provision of the database utilizes the detection of scattered light from each pattern structure (by bright field or high resolution dark field detection mode) in a plurality of structures, and analyzes the data indicative thereof. This analysis procedure includes modeling an image of the pattern in the shape of a two or triad array of discrete scattering, and determining the intensity distribution of the scattered light from the patterned model within the maximum condensing solid angle.

According to yet another aspect of the present invention, there is provided an optical inspection device for inspecting a patterned article, the apparatus comprising:

A dimming system for dimming an area on the product;

A condensing system for condensing light components scattered from the illuminated region of a predetermined condensing angle field located outside the solid angle of propagation of the specularly reflected light components;

A filter operable to separate the light component propagating to the predetermined solid segment of the predetermined solid field in the optical path of the scattered and focused light component;

A detection system having a detection unit for receiving the filtered portion of the scattered and focused light and generating data indicative thereof; And

A control unit for operating the filter in an optical path of scattered and focused light, in accordance with predetermined data representative of at least one solid-angle segment of light propagated from the pattern in the illuminated area.

Certain data may have been previously acquired by the same optical inspection apparatus or both by using a bright field or dark field detection mode.

Preferably, the filter is in the shape of a mask assembly consisting of a plurality of different masks which propagate to at least one predetermined solid angle segment of the maximum condensing angle for filtering light components or correspond to a predetermined pattern structure. Such mask assembly may optionally be operated such that an optional mask is positioned in the optical path of the focused light. The mask may be LCD, mechanical or the like. The filter is preferably located within the Fourier plane (conjugate plane).

The present invention also provides, according to another aspect, an optical inspection device for inspecting a patterned article, the apparatus comprising:

(a) a dimming system for dimming an area on a product;

(b) a first condensing mechanism for condensing light components returned from the dimmed region, and a second condensing mechanism for condensing light components scattered from the dimming region in a predetermined condensing angle field located outside the solid angle of the specularly reflected light component; Condensing system;

(c) a filter selectively operable to separate the light component propagating in the optical path of the scattered and focused light component to a predetermined solid segment of the predetermined focusing angle field;

(d) a first detection unit for detecting the light component collected by the first light collecting mechanism and generating data indicative thereof, and a second detection unit for detecting the filtered portion of the scattered and focused light and generating data indicative thereof A detection system having a; And

(e) a control unit responsive to the data generated by the first detector for analyzing the data to operate the filter.

According to still another aspect of the present invention, there is provided a control unit for use in a system for optical inspection of a pattern structure with a predetermined maximum solid angle of light scattered from the inspection target structure, wherein the predetermined maximum condensing solid angle propagates light that is specularly reflected from the structure. Located outside the solid angle of the control unit, wherein the control unit includes a memory in which a database is stored, the database comprising a plurality of data records comprising a pattern structure to be inspected, wherein the data records are respectively compared to the other solid angle segments of the maximum condensing solid angle; Data representing one solid angle segment representing at least a predetermined maximum condensing solid angle corresponding to the direction of propagation of light components scattered from the corresponding pattern structure, wherein the contribution of light scattered from the pattern is substantially small.

More specifically, the present invention is used to inspect a semiconductor wafer (which constitutes a patterned product or pattern structure in view of the structure manufactured by one of the manufacturing steps applied to the same product), and thus, the application thereof. It demonstrates.

Referring to FIG. 1, one possible embodiment of the optical inspection apparatus 1 according to the invention associated with the wafer W to be inspected is shown. The apparatus 1 includes a light control system 2 and a light condensing / detecting system. The latter consists of a first condensing optics associated with the detection unit 4 operating in the bright field image mode, and a second condensing optics (not shown) associated with the detection unit 6 operating respectively in the dark field image mode. . In this embodiment, four such detection units 6 are provided, each associated with a corresponding focusing optics. The output circuit (not shown) of the detection units 4, 6 is associated with the control unit 7. Usually, the control unit 7 is a computer system which is suitably equipped with hardware and operated by appropriate software to successfully analyze the data reached from the detection unit, which will be described in more detail below.

The dimming system 2 usually includes a light source 8 such as a laser that irradiates a beam of light at 10 and a light directing optical mechanism 12. In the case of using a laser as a light source, the optical apparatus 12 includes a suitable scanning means 14 (e.g., an acoustooptical device) mounted on the optical path of the beam 10 and a focusing optical mechanism 15 (e.g., an objective lens). ) When operated by laser dimming, a high numerical aperture objective lens is used to provide a high resolution of the acquired image. While the wafer W is supported on a translation stage to move along the Y-axis, the scanning means 14 moves the beam 10 in the scanning direction (eg, the X-axis). . The scanning means 14 and the focusing optics 15 work together to focus the beam onto the scan line S (which constitutes the illuminated area) on the surface of the wafer. Alternatively, other suitable scanning means such as a rotating mirror can be used. Provision of the scanning means is optional and may use a "biscanning" beam for the same purpose, ie for dimming the line on the wafer surface. Here, the propagation of light is schematically shown in a single line in order to facilitate the depiction of the main parts of the apparatus 1.

In this embodiment of FIG. 1, the dimming system 2 provides a vertical incident beam 10 on the wafer W. As shown in FIG. At the same time, the light directing optics 12 includes a beam splitter 16 for separating the incident and reflected light components. However, it should be noted that the incident beam may be directed onto the surface of the wafer at a predetermined angle of incidence. It should also be noted that, in general, bright-field based systems may utilize both the same dimming path or its own dimming path used for dark field images.

As described above, in the present embodiment, the detection unit 4 operates in the bright field image mode, that is, condenses the light component 10A specularly reflected from the scan line S, where the specularly reflected light It propagates within the solid angle of. The detection unit 4 comprises a suitable detector, such as PMT, CC, or pin-diode, which can collect and receive specularly reflected light and generate data representing an image of the scan line S.

Each detection unit 6 operates in a dark field image mode, ie focuses the rising difference at the azimuth and height different from where the most specular reflection of the light component 10B scattered from the surface of the wafer occurs.

Generally speaking, as described below, the condensing optics of the detection unit 6 have a predetermined maximum condensing solid angle (constituting the condensing angle field). The detection unit 4 enables the analysis of the image and the determination of at least one solid angle segment of the condensing angle field, which is used for obtaining an image of the pattern structure, and used for detecting light with the detection unit 6. . At least one solid angle segment is light whose dispersion of light returned from the pattern is small compared to that of the other solid angle segments of the focusing angle field.

Thus, in the embodiment of FIG. 1, even if the detection unit 4 is operated in the bright field detection mode, image acquisition and analysis are possible, and thus can be used for high-resolution dark field detection. For imaging and inspection purposes, a high-resolution dark field detector can be used.

The invention may typically utilize a detector of an auto-focus system installed in an optical inspection apparatus as a bright-field detector for imaging of a pattern. In FIG. 2, a conventionally designed, optical inspection apparatus 100 comprising a dimming system 102, an auto-focus system 103 operating in a bright field detection mode, and four dark-field based detection units 6. Is shown. The detector 6 and that of the auto-focus system 103 (not shown) are connected with the control unit 7. The detector of the system enables the imaging of the illuminated area on the product, which image is analyzed by the control unit to determine the condensing angle of the maximum condensing angle field, which is used for detecting light with the detection unit 6. .

Since the four detection units 6 in the two embodiments of FIGS. 1 and 2 have a similar configuration, only one main configuration is described with reference to FIG. 3. The detection unit 6 includes the condensing system 18 and the detector 20, and may make its configuration and operation into a suitable kind, such as a known PMT.

In a preferred embodiment, the condensing system 18 comprises, after the optical system 30, a condensing optics 22 (which constitutes the second condensing optics), an interference (imaging) fiber bundle 26, image propagation means. And mask assembly 28 (constituting the filter). Although not an essential element, in a preferred embodiment, it includes a polarizer inserted in the optical path of the optical instrument 22. The principle of operation of the interference fiber is known and need not be described in detail except as mentioned below. The interfering fiber bundle is used for the transmission of the image. The relative position of the individual fibers in the interfering fiber bundle is maintained, and due to this fact the fiber bundle provides a desired linear match between the input and output signals.

Therefore, the incident light 10 affects the scan line S. FIG. As shown in FIG. 1, the specularly reflected light component 10A is collected by the optical instrument 3a and received by the detection unit 4. Referring to FIG. 3, light components 10B scattered from patterns and abnormal shapes which may often occur in the scan line S on the surface of the wafer are propagated to the optical instrument 22. Light components 10B scattered from the scan line and propagated at a predetermined maximum solid angle (condensing angle field) are collected by the optical instrument 22 and directed to the polarizer 24. The polarizer 24 is operated with a suitable driver (not shown) to move in such a way that it is located in both the optical paths of the light guaranteed from the condensing optics 22. In addition, when installed in the optical path, the polarizer 24 can be rotated in such a manner that the orientation (called polarization) of the desired transmission plane is changed. The transposition of the polarizer 24 affects the light passing through it. Since the pattern may have a high degree of polarization, represented by "polarized noise," the polarizer's transposition affects the signal-to-noise ratio of the light component propagating towards the detector. In addition, the running mode may be applied to a predetermined kind of product prior to the inspection. This enables a predetermined polarization signal in the light component returned from the desired pattern, and thus, detects a difference (defect) by changing the orientation of the polarizer 24.

The condensing optics 22 is then designed to form an angular image of the scan line S which is transmitted to the detector 20 via the fiber bundle 26. As shown in FIG. 4, the condensing optics 22 includes first and second assemblies 32, 34, and a slit 36 located at the front focal plane P of the lens assembly 34. Light propagation through the lens assemblies 32, 34 defines the optical axis OA. The lens assembly 32 is positioned to focus light components scattered from the scan line S and propagating at a predetermined solid angle. The dimension of the first glass surface 32A of the lens assembly defines the value of the solid angle of condensing of the optical instrument 22, which is the maximum condensing angle of the entire detection unit 6 (called the condensing angle field). The lens assembly 32 is designed to form the actual image S 'of the scan line S in the image plane IP. The image plane IP substantially coincides with the front focal plane P of the lens mechanism 34. The slit 36 is shaped and dimensioned similar to that of the actual image S ', and installed at the location of the desired image. The lens assembly 34 is formed from the actual image S 'of the scan line S and the angled image S ". Each point of the angled image S" is any point of the scan line S. Formed by light components propagated from and propagated into segments of the maximum condensing angle. This segment of each maximum condensing angle is irradiated to the corresponding point (fiber) on the input surface 26A of the fiber bundle 26 and propagated by the fiber bundle at the same angle assured from the corresponding point on the output surface of the fiber bundle 26. .

The slit 36 installation is based on: Due to the internally generated reflections, the dimming unit forms an additional undesirable image (called "ghost") of the scan line on the surface of the wafer. Undesirable images will also cause light propagation back to the focusing optics. The slit 36 thus designed and positioned is the passage of condensed light represented by the actual image S 'of the scan line S with the lens assembly 34 and the passage of light other than that which is combined into the image S'. To block.

The surface 26A defined by the front end (corresponding to the light propagation) of the fiber bundle 26 is located in the focal plane FP of the entire condensing optics 22, thereby allowing the entrance pupil of the system to enter. Defines the most preferred position for. The design and location of such condensing optics 22 relative to the fiber bundle 26 allows the resolution of the condensing system 18 to be significantly increased and the fiber bundle operation to be optimized. The diameter of the fiber bundle 26 is preferably slightly exceeded so that all condensing beams (eg, solid angle segments of the maximum condensing angle) defined by the cone of the maximum condensing angle enter the fiber bundle 26. To do that.

Thus, the condensing optics 22 forms an angular image S ″ of the scan line S and directs the focused light representing the image to the fiber bundle 26. Each of the images S ″ The point appears at an angle of propagation (in the condensing angle field) of condensing components scattered from any point on the scan line S. The light component at each angle is irradiated at a corresponding point (fiber) on the input surface 26A of the fiber bundle 26 and is followed by the fiber bundle, resulting in a corresponding point on the output surface of the fiber bundle. Propagated at an angle. This technique ensures the same angular field for all points along the scan line S and for the detection unit 6, thus comparing the data generated by all the detection units 6, of the differently indicated signals that can be detected. Increase the signal to noise ratio.

According to FIG. 5A, the mask assembly 28 is composed of a plurality of differently designed masks formed on a conventional disk-shaped opaque plate 40. For example, it may be a metal disk having holes or a chromium-containing glass mask. The disk 40 is rotated by a motor for rotation in the plane perpendicular to the long axis of the fiber bundle, to selectively position the desired mask in the optical path of the propagating light, which is not specifically shown here.

In the present invention of FIG. 5A, a plurality of different masks are arranged in a periodic pattern formed by transparent or opaque regions, and extend along the peripheral region of the disk 40. Each pattern region is formed by another combination of locally adjacent transparent and opaque regions and represents one mask from a plurality of different masks formed by other regions of the pattern. One transparent / opaque region, combined with the neighboring region on the left and the neighboring region on the right, may form two different regions, respectively. By rotating the disk about these axes, different patterns of area (masks) are located in the optical path of the condensed light colliding on the disk.

5B shows a mask assembly 128 having several designed differently compared to the mask assembly 28 of FIG. 5A. Here, the mask assembly is composed of a plurality of differently designed masks, each of which is designed in the present embodiment as 128a to 128g, and is formed on a conventional disk-shaped opaque plate 40. In the operating position of the mask assembly 128, one of the masks 128a-128g is located in the optical path of light from the fiber bundle 26. Each of the masks 128a and 128b represents an aperture defining a certain condensing solid angle smaller than the maximum condensing angle defined by the first glass surface 32A of the lens mechanism 32. By positioning one of the masks 128a, 128e in the optical path of light from the fiber bundle 26, one predetermined solid angle segment of the maximum condensing angle may be picked up. Masks 128b, 128c, 128d, 128f, and 128g have different patterns formed in regions R _T and R _B that transmit and block light collisions on the mask, respectively. Each of these patterns can cut off or pick up one or more solid segments at the operating position (eg, located in the optical path of light from the fiber bundle).

Typically, each mask propagates at least one propagation region, corresponding to a condensation corresponding to at least one solid-angle segment of the maximum condensing angle (field of condensing angle). In the operating position of the mask assembly 28, or 128, one of the masks is installed in the optical path of the light from the fiber bundle 26, ie, the conjugated surface of the wafer face, corresponding to the solid angle segment with respect to the detector 20. It propagates the light component.

The supply of a mask assembly having a plurality of different masks selectively changes the condensing angle of the entire detection unit 6, thereby providing a custom condensing. It is important that such a mask assembly provides a solid angle change simultaneously along two mutually perpendicular axes, represented by A ₁ and A ₂ , which resemble altitude and azimuth scattering angle.

For the purposes of the present invention, light components scattered from the pattern appear as "noise" in which the detected signal is filtered, while "light defects" in which light components scattered from the anomaly are detected. Knowledge of the wafer pattern is necessary to determine that the solid angle segment is filtered with light propagating towards the detector 20 or that the solid angle segment should be selected as the primary produced condensation. This pattern may be represented by an auto-focus system (103 in FIG. 2) or a high-resolution dark field detector in the bright field image of the scan line obtained from the detection unit (4 in FIG. 1).

Therefore, the data representation of the image is appropriately generated and propagated to the control unit 7 which analyzes and "models" this data, thereby constructing a simulated image of the pattern. This is shown in Figures 6a and 6b. FIG. 6 shows the actual image I _pat of the area illuminated in a nearly perfect periodic two-dimensional pattern shape of the elementary cell element CE. Modeling of this pattern structure involves replacing all elements CE in the repeating cell with point-scattering objects. FIG. 6B shows the model SM of the structure I _pat formed by discontinuous scattering DS.

The steps in image modeling include the following steps:

(1) a cell structure consisting of correcting the matrix of the image Ipat itself, without exhibiting zero variation, and determining a first maximum value corresponding to the base cell dimension for the correction variation in the two external dimensions recognition.

(2) While the elementary cell structure may be selected from a list of known cell structures such as FCC, BCC, a cell structure in which an image is constructed using the structure and dimensions determined in step (1), and the result corresponds to the original image. recognition. The structure giving the maximum corresponding value is the desired cell structure.

(3) Infer the point scattering for each elementary element CE.

The result of the image modeling step is a two-dimensional discrete scattering function (S _ij ), which represents a scattering pattern from the wafer surface that is proportional to the scattering intensity.

The data representation of the distance between two repeated discrete scatters along the X and Y axes (e.g., " pitch ") of the known structure is previously stored and stored in the memory of the control unit. This is shown in FIGS. 7A-7E, respectively, showing four embodiments of known structures such as FCC, BCC, diagonal and line structures. Here, Δx and Δy are the structural pitches along the X- and Y-axes.

In order to consider the actual scattering pattern, the actual shape and size of the dimming function should be irradiated on the scattering surface. That is, a step of periodic scattering calculation is performed. For this purpose, the dimming function is set, and the discontinuous sum of scattering from each element colliding with this function is performed.

(A) Dimming function setting

By selecting the dimming function model as a model of spot incidence with limited diffraction on the wafer, a two-dimensional discontinuous function L _ij is obtained, which represents the incident intensity on each discontinuous portion on the wafer surface. The discontinuous dimming function consists of a truncated Gaussian beam with various parameters fixed such as spot size, angle of incidence, image distortion along the spot size, and position along the scan line. The spot on the wafer is the Fourier transform of the truncated Gaussian, characterized by the following parameters.

Spot size at 1 / e ² for both spindles (scan and cross scan);

-Apodization width and position for maximum intensity of spot;

Spot telecentricity simulated with a linear phase difference along one or two spot principal axes;

Normal phase distortion across the spot; And

The length of the scan line

The result of the foregoing step is a two-dimensional discontinuous function (I _ij ) which represents the incident intensity on each discrete portion on the wafer surface.

8 shows an embodiment of a simple Gaussian spot. Here, the distribution RMS is 10 μm, thus modeling a spot size of 20 μm and the total intensity is normalized to one.

Next, the result of the scattering intensity of each point on which the object is scattered on the plane is calculated by the respective product of the scattering structure S _ij and the dimming function I _ij .

(B) Scattering scattered scattered point scatters on the surface at all angles in the condensing angle field is summed and shown in FIG. 9 using the shape of the scattering problem.

Using the vector wave formula, the resulting wavelength distribution is computed as each of the many spheres surrounding the reflection pattern by the amplitude and phase sum of the surface field dispersion. As is known, the conventional vector form of far field electric field by small dipoles is given by

Where r is a vector defining the position of the detector; P is a vector defining the vibration orientation of the dipole; And k is the wave number (| k | = 2π / λ, where λ is the wavelength of the beam). Note that instead of point scattering objects, formal elements can be considered.

For a given incident wavelength polarization p ₀ , it is assumed that the scaling element does not exist in such a state that p = p ₀ .

The sum of the fields is as follows:

here,

And R is the distance of the detector.

The resulting intensity distribution in the sphere of the detector is given by

Simulated with a simulated dark field scattering pattern, this result usually consists of a plurality of strong lobes located at an angle. This simulation model pattern can be compared with the imaging of the actual scattering intensity along the condensing angle field. 10A and 10B show what is obtained by simulation of scattering from the wafer structure and the actual pattern seen with the imaging fibers, respectively. Thus, this demonstrates that by blocking the rising "background" intensity from the periodic pattern seen as a strong lobe in the simulation imaging plot, the signal-to-noise ratio in the detected signal can be significantly increased.

Returning to FIGS. 5A and 5B, an effective dark field mask can be found based on an optimal algorithm that minimizes structural scattering intensity while maximizing the angular vacancy of the condensing angle field. This effective mask can be used while scanning the wafer.

It is important to note that most patterned products are made in series. Thus, each product type or structure obtained with one of the manufacturing steps applied to the same product can already be imaged to determine the solid angle segment corresponding to the propagation of the scattered light from the pattern within the maximum condensing angle of a given optical instrument. Based on this data, a library can be prepared and stored in the memory of the control unit. In this case, it is not necessary to actually apply a certain detection mode (bright field or dark field detection) for acquiring and analyzing the image of the inspection target structure, but operating the mask assembly (filter) by using some reference material from the library. . The other kind of product is different from the pattern structure of the same wafer or the like, but not the wafer after the step of different manufacturing with the different patterned top layer.

Although not shown in detail, it should be noted that the mask assembly 28, or 128, can be accommodated in the optical path of the focused light upstream of the fiber 26. In this case, the mask assembly should be located at the focal plane FP of the condensing optics 22, and the fiber bundle 26 should or should not be image fiber.

It should also be noted that such mechanical filters may be replaced with known micro electro-mechanical structures (MEMS) or programmable liquid crystal displays (LCDs). LCD segments are light valves, where light is transmitted or reflected in open mode (depending on the type of liquid crystal material used therein), and when light is blocked in close mode. A spatial filter in the form of a radio wave LCD is disclosed by way of example in US Pat. No. 5,276,498.

If a reflective LCD or MEMS is used as the mask assembly, it will be appreciated that the fiber bundle is positioned to receive the light collected from the filter (selection mask). This fiber bundle does not require an imaging form.

Referring again to FIG. 3, the optical system 30 is designed to irradiate light collected by the mask assembly 28, or 128 throughout the sensing surface of the detector 20. For example, the system 30 may be a telecentric imaging system, where the sensing surface of the detector 20 is located at the location of the trans pupil of the system. The principle of telecentric imaging optics is known, except that such optics do not require long-range injection magnification changes and maintain the same magnification of the image over a wide range along the optical axis of the system. It does not need to be described in detail. Provision of an optical apparatus for irradiating the light collected by the mask assembly to the entire sensing surface is configured as follows. Typically, the sensing surface of the detector has a non-uniform sensing distribution. If spatially separated light cones (by masks) were irradiated to separate areas on the corresponding sensing surface, the detection difference of the input signal generated by the detector would be related to the detection difference of these areas and to the difference of scattered light. There is no number. To avoid this undesired dependence between the output signal and the selected focusing angle, the optics 30 provide dimming of the entire sensing surface by each one of the focused light components (selected angles).

Those skilled in the art can make various modifications and changes such as the preferred embodiments of the above-described invention without departing from the scope of the invention by the appended claims. For example, the light collecting / detecting assembly for imaging purposes may be operated in a high-resolution bright or dark field detection mode. The fiber bundles may be replaced with other suitable image conversion means, such as radio optics, which are conventional multilens systems for maintaining a certain number of apertures of light propagation. The mask assembly may comprise a plurality of existing separate masks of conventional disks, rather than consist of a periodic pattern of transmission and opaque regions, and may be installed both before and after the image transmission means.

Claims (21)

(Iii) dimming an area on the product with incident light to form light returned from the dimming area; (Ii) acquiring an image of the dimming area and generating data indicative thereof; (Iii) analyzing the generated data and determining the intensity distribution of the light components scattered from the pattern of the dimming area within a predetermined condensing angle field outside the solid angle of propagation of specularly reflected light; step; And (Iii) based on the determined distribution, by filtering the light collected in the predetermined condensing angle field, the light component scattered from the dimming area and propagated to at least one predetermined solid angle segment of the predetermined condensing angle field; Focusing and guiding the collected light components to a detection device. The method of claim 1, wherein analyzing the generated data comprises: Constructing a model of the pattern structure of the illuminated area to obtain a discrete array representing the scattered pattern; Simulating a dark-field scattering pattern within the predetermined focusing angle field; And And determining the at least one solid angle segment as a position where the background intensity becomes smaller than other segments of the predetermined solid angle by determining a rising background intensity from the periodic pattern based on the simulation result. Optical inspection method of the patterned product made into. The method of claim 2, wherein constructing the model comprises: Determining a dimension of the elementary cell that is an element of the pattern repeated several times the two lateral dimensions; Determining a cell structure in the pattern; And And assuming scattering points for each base cell. The method of claim 1, wherein the acquired image is obtained in a bright field detection mode. The method of claim 1, wherein the acquired image is obtained in a dark field image detection mode. The process of claim 1 wherein the pattern product processing on a manufacturing line is a stream of similar pattern products, wherein steps (ii) and (iii) are performed only for the first pattern product in the stream. Optical inspection method of a patterned product. A method of inspecting a pattern structure at a predetermined maximum condensing solid angle of light scattered from the inspection target product, which is outside the solid angle of the propagation of light specularly reflected from the pattern structure, Corresponding to a plurality of pattern structures including the pattern structure to be inspected, wherein each of the solid segments is scattered from a corresponding pattern structure in which the distribution of light scattered from the pattern is substantially smaller than other solid segments of the solid angle of the condensing. Providing a database comprising data representing a plurality of solid angle segments of light propagation within the predetermined maximum condensing solid angle, respectively corresponding to the propagation direction of the light component to be; And Selecting the data representing the pattern structure from the database, the data filtering the light component propagating from the total light collected at the predetermined maximum condensing angle to a corresponding solid angle segment to detect the light component. A method for inspecting a pattern structure, comprising the step of being used. An optical inspection device for inspecting a patterned product, A dimming system for dimming an area on said product; A condensing system for condensing light components scattered from said dimmed regions of a predetermined condensing angle field located outside the solid angle of propagation of specularly reflected light components; A filter operable to separate light components propagating in a predetermined solid segment of the predetermined solid field in the optical path of scattered and focused light components; A detection system comprising a detection unit for receiving the scattered, focused, filtered portion of light to produce data indicative thereof; And And a control unit for operating the filter in an optical path of scattered and focused light, in accordance with predetermined data representative of at least one solid-angle segment of light propagated from the pattern in the dimming area. 9. The system of claim 8, further comprising a light collecting / detecting system for generating data representing said image to be analyzed by a control unit for obtaining an image of the illuminated area and determining said predetermined data. Optical inspection device. 10. An optical inspection apparatus according to claim 9, wherein said acquired image is obtained in a bright field detection mode. 10. An optical inspection apparatus according to claim 9, wherein said acquired image is acquired in a dark field detection mode. An optical inspection device for inspecting a patterned product, (a) a dimming system for dimming an area on said product; (b) a first condensing mechanism for condensing light components returned from the dimmed region, and a second condensing mechanism for condensing light components scattered from the dimming region at a predetermined condensing angle field located outside the solid angle of the specularly reflected light component; A condensing system comprising a; (c) a filter selectively operable to separate the light component propagating in the optical path of the scattered and focused light component to a predetermined solid segment of the predetermined focusing angle field; (d) a first detecting unit for detecting the light component collected by the first light collecting mechanism and generating data indicative thereof, and a second portion for detecting and generating the filtered portion of the scattered and focused light; A detection system comprising two detection units; And (e) a control unit responsive to data generated by the first detector for analyzing the data to operate the filter. 13. The apparatus of claim 12, wherein the filter filters each of at least one solid angle segment of the condensing angle field and allows propagation to a second detection unit to block another solid angle segment of the condensing angle peel. Optical inspection device comprising two different masks. 14. An optical inspection apparatus according to claim 13, wherein the plurality of different masks exhibit a continuous pattern formed by transparent and opaque regions and extending along the peripheral region of the disc-shaped plate. 14. An optical inspection apparatus according to claim 13, wherein said other masks are arranged in a spaced apart relationship along the circumferential region of the disc-shaped plate. The method of claim 12, wherein the analysis of the data generated by the first detection unit is continuous such that the solid angle segment of the at least one condensation is determined such that the background intensity is smaller than other segments of the predetermined condensing angle field. Determining a background intensity rising from the typical pattern, wherein the filter is operative to filter from totally scattered and focused light propagating to the at least one solid segment. The method of claim 16, wherein the control unit is configured by the first unit to receive data representing an image of the pattern of the illuminated area, and to control the pattern structure of the illuminated area in a discrete array shape representing a scattering pattern. And a process facility for constructing a model and simulating a darkfield scattering pattern within said predetermined condensing angle field. 13. The optical as claimed in claim 12, further comprising at least one additional detection unit for detecting light scattered from the dimming area, wherein the at least two detection units detect a light portion focused at different azimuth angles. Inspection device. 13. The optical inspection apparatus according to claim 12, wherein the condensing and detection of the light returned from the dimming area uses a bright field detection mode. 10. The optical inspection apparatus according to claim 9, wherein the condensing and detection of the light returned from the dimming area uses a dark field detection mode. A control unit for use in a system for optical inspection of a pattern structure by a predetermined maximum condensing solid angle of light scattered from a structure to be inspected, which is located outside the propagation of light specularly reflected from the structure, The control unit includes a memory in which a database including a plurality of data records including the pattern structure to be inspected is stored; The data records each contain at least the predetermined maximum condensing solid angle corresponding to the direction of propagation of light components scattered from the corresponding pattern structure, wherein the distribution of light scattered from the pattern is substantially smaller than the other solid segments of the maximum condensing solid angle. And a data unit indicating one solid angle segment to be shown.